Model Construction And Optimization Strategy Research Of Ridesourcing And Traditional Taxi Selection

The existing taxi modes are mainly divided into traditional taxis and app-based ridesourcing.As two similar yet different modes of transportation,the imbalanced development of the two modes can increase the probability of traffic congestion and lead to the inefficient use of vehicles.Therefore,this study focuses on the two travel modes,takes ridesourcing and traditional taxi choices with and without time constraints as the key technology,reveals the influencing mechanism as the theoretical basis,and propose quantitative optimization measures.The aim is to promote the balanced development of the two taxi modes and fully utilize their business characteristics to enhance overall service efficiency.First,the differences in travel characteristics between ridesourcing and traditional taxis were compared.Payment amount differences under different time periods for the same origin/destination(OD)were analyzed.Furthermore,differences in the spatial and temporal aspects of choice behavior were compared.The results showed that the ridesourcing are higher in overall travel characteristic mean and dispersion than traditional taxis.However,traditional taxis are more efficient and accurate in their route choices.Ridesourcing fares are higher than traditional taxis in most time periods,and a larger price difference during peak hours.The spatial and temporal distribution of the two modes also exhibited distinct regional characteristics,and their competitive situation also had distinct regional characteristics.Second,the city was divided into six functional areas based on hierarchical clustering algorithms.The ridesourcing and traditional taxi OD data were then combined with functional areas to identify the travel purposes of single trips.According to the travel characteristics,travel purposes were divided into two categories: trips with and without time constraints.Trips with time constraints were considered as having higher reliability requirements.Analysis of the difference in characteristics between the two types of trips showed that trips with short distances,short in-vehicle time and lower payment amounts had higher reliability.The change in ridesourcing characteristics is more affected by the with or without of time constraints,and changes in characteristics within different ranges can lead to changes in travel preferences.Third,to understand the influence mechanism of travel characteristics on choice behavior preferences,this study constructed two models of ridesourcing and traditional taxis with or without time constraints,using random forest algorithms and partial dependence plots to reveal the feature importance and non-linear relationship of travel characteristics on choice probability.The study found that during regular rest time periods,for trips with lower payment amounts,with time constraints and originating from areas containing train station POIs,people are more likely to choose traditional taxis.On the other hand,for trips with higher payment amounts,shorter passenger distances,without time constraints and longer in-vehicle times,people are more likely to choose ridesourcing.In terms of reliability,traditional taxis have higher reliability when there are more populations or POIs in the origin area,while ridesourcing have overall higher reliability.The non-linear relationship between transportation facility characteristics and choice preferences also indicated the inhibitory and promotive effects between different transportation modes.Finally,price adjustment optimization measures and comprehensive adjustment optimization measures were proposed to transfer trips between the two modes.Different optimization measures were selected for trips in different scope and categories to achieve more effective optimization.The results showed that using price adjustments within the sensitive range of average trip payments achieved better results,while adjusting prices again after average trip payments approach or reach their threshold had weak effects.Comprehensive adjustment could significantly enhance the optimization effect.Overall,ridesourcing users exhibited higher loyalty.Fully utilizing the respective business characteristics of ridesourcing and traditional taxis can enhance overall service efficiency.